Video Screencast Help

Patching Solaris 10 servers with CVM Cluster

Created: 11 Sep 2012 | 4 comments

Hello,

I've many servers for patching/minor upgrade to U9, all the servers are part of CVM cluster + root disks are part of either SVM or ZFS . I would like to seek information from experts on how to go about this patching/Upgrade.

If someone has any procedure for this then please share it with me.

Thank you/Nilesh

Discussion Filed Under:

Comments 4 CommentsJump to latest comment

Venkata Reddy Chappavarapu's picture

Did you check the SFCFSHA install guide? It has a section on how to upgrade OS. For example, If you have SFCFSHA 6.0 installed you can follow the procedure described below:

https://sort.symantec.com/public/documents/sfha/6....

For complete guide you can refer to, http://www.symantec.com/business/support/resources...

Thanks,

Venkat

Venkata Reddy Chappavarapu,

Sr. Manager,

Information Availability Group (VCS),

Symantec Corporation

===========================

PS: If you are happy with the answer provided, please mark the post as solution. You

Nilesh.Joshi's picture

Thanks Venkat, however not of much help.

Anyone in community who can guide me on this?

Requirement -

Patch/Upgrade CVM nodes with root disk either on SVM or ZFS. Please share the process if anyone has it handy.

Thank you/Nilesh

 

 

Leed.Engineer's picture

Hello Nilesh ,

     First step to do is to break the root mirror in case there is any unstable patch in the patch set you are going to install , using the below commands :

metadetach d0 d20 ---> submirror

metadb -d <submirror_disk>

metaroot <main_disk,c#t#d#s0> , Now you have golden copy of data in case of anything wrong happens.

After that you have one of two ways:

Live Upgrade : and you will need to switch over all the groups to the other node then offline the remaining parallel SGs in the current node. Then , start the upgrade on this node and vice versa.

 

Offline Upgrade :

In this case you will need to offline all the SGs on all Systems and kick off the upgrade one the nodes.

 

In both cases you will need to do the root mirror break.

 

Regards

Waleed Badr

 

 

krishna.karthikeya@gmail.com's picture

Hi Nilesh,

I would like to express below procedure would may help to patching on solaris box if CVM running....

A) UFS + CVM
    
    1. disable vcs script on all running host --- mv /etc/rc3.d/S99vcs /etc/rc3.d/s99vcs
    2. Detach rootmirror device from main device -- metadetach <main> <submirror>

    3. mount submirror device under /mnt(any temp mount point) and change the root and var entires accordinly  and execute metaroot -k /mnt/etc/system -v /mnt/etc/vfstab <submirror>

   4. Install bookblock on submirror

   5. shutdown and reboot test using submirror

   6. once you feel that rootmirror working fine.. reboot with help of rootdevice in single user mode for patching

   7. Once u hv patched successfully, Reboot with help of rootdevice and start the cluster manually..
      Preferred to start cluster first in master node as per details before reboot.

      In Master node -- start the seeding first  -- gabconfic -xc
      # hastatus -sum and vxdctl -c mode
      Finally start all slave nodes which are in part of ur cluster.

 

Backout : boot submirror and start the cluster

####################################################################

B -- ZFS and CVM

Same CVM procedure you can use.

In case zfs.. take the snapshots for root and var
 

go patching...

 

Backout : boot failsafe mode and rollback root and var

take the reboot

 

Above procedure will may help you...

Thanks 

 

Krishna...